A higher-dimensional Chevalley restriction theorem for orthogonal groups (Q6103494)

Let \(G\) be a reductive group over an algebraically closed field \(\mathbb{K}\) with Lie algebra \(\mathfrak{g}\), and denote by \(\mathfrak{C}^d_{\mathfrak{g}}\) the corresponding \emph{commuting scheme}, whose underlying variety \(\mathfrak{C}^d_{\mathfrak{g},\mathrm{red}}\) consists of \(d\)-tuples \((x_1,\dots,x_d)\) of elements of \(\mathfrak{g}\) with \([x_i,x_j]=0\) for all \(1\le i<j\le d\). The embedding of the Cartan subalgebra \(\mathfrak{t}\) into \(\mathfrak{g}\) induces a morphism \(\Phi:\mathfrak{t}^d/\negmedspace/W\to \mathfrak{C}^d_{\mathfrak{g}}/\negmedspace/G\), where \(W\) is the Weyl group and \(\mathfrak{C}^d_{\mathfrak{g}}/\negmedspace/G\) is the categorical quotient of \(\mathfrak{C}^d_{\mathfrak{g}}\) by \(G\). In connection to their study of the Hitchin morphism from the moduli stack of principal \(G\)-Higgs bundles over a proper smooth variety, Chen and Ngo conjectured that \(\Phi\) is an isomorphism for any reductive group \(G\). The special case \(d=1\) of this conjecture is the Chevalley restriction theorem. The special case \(\mathrm{char}(\mathbb{K})=0\) and \(G=GL_n(\mathbb{K})\) was proved independently by Domokos and Vaccarino, whereas the case \(G=Sp_n(\mathbb{K})\) was proved by Chen and Ngo. The main result of the present paper is that \(\Phi\) is an isomorphism when \(\mathrm{char}(\mathbb{K})=0\) and \(G\) is the orthogonal group \(O_n(\mathbb{K})\) or the special orthogonal group \(SO_n(\mathbb{K})\). Moreover, when \(\mathrm{char}(\mathbb{K})>2\) and \(G\) is \(O_n(\mathbb{K})\) or \(SO_n(\mathbb{K})\), the weaker statement \(\mathfrak{t}^d/\negmedspace/W\cong \mathfrak{C}^d_{\mathfrak{g},\mathrm{red}}/\negmedspace/G\) is proved. We get as a corollary that if \(\mathrm{char}(\mathbb{K})=0\) and \(G=O_n(\mathbb{K})\) or \(SO_n(\mathbb{K})\), then the scheme \(\mathfrak{C}^d_{\mathfrak{g}}/\negmedspace/G\) is integral (i.e. reduced and irreducible) and normal. Note that it is a well-known open question whether the commuting scheme itself is reduced.